Content uploaded by Jonathan P. Folland
Author content
All content in this area was uploaded by Jonathan P. Folland
Content may be subject to copyright.
Strength training: Isometric training at a range of joint angles versus
dynamic training
JONATHAN P. FOLLAND
1
, KATE HAWKER
2
, BEN LEACH
2
, TOM LITTLE
2
,&
DAVID A. JONES
2
1
School of Sport and Exercise Sciences, Loughborough University, Loughborough, and
2
School of Sport and Exercise Sciences,
University of Birmingham, Birmingham, UK
(Accepted 9 October 2004)
Abstract
Strength training with isometric contractions produces large but highly angle-specific adaptations. To contrast the contractile
mode of isometric versus dynamic training, but diminish the strong angle specificity effect, we compared the strength gains
produced by isometric training at four joint angles with conventional dynamic training. Thirty-three recreationally active
healthy males aged 18 – 30 years completed 9 weeks of strength training of the quadriceps muscle group three times per
week. An intra-individual design was adopted: one leg performed purely isometric training at each of four joint angles
(isometrically trained leg); the other leg performed conventional dynamic training, lifting and lowering (dynamically trained
leg). Both legs trained at similar relative loads for the same duration. The quadriceps strength of each leg was measured
isometrically (at four angles) and isokinetically (at three velocities) pre and post training. After 9 weeks of training, the
increase in isokinetic strength was similar in both legs (pooled data from three velocities: dynamically trained leg, 10.7%;
isometrically trained leg, 10.5%). Isometric strength increases were significantly greater for the isometrically trained leg
(pooled data from four angles: dynamically trained leg, 13.1%; isometrically trained leg, 18.0%). This may have been due to
the greater absolute torque involved with isometric training or a residual angle specificity effect despite the isometric training
being divided over four angles.
Keywords: Muscle strength, isometric, dynamic, isokinetic, resistance training
Introduction
The most effective means of increasing strength by
high resistance training remains unknown, despite
the obvious importance of this knowledge for
athletic training and rehabilitation. We have re-
cently examined the influence of fatigue in
resistance training (Folland, Irish, Roberts, Tarr,
& Jones, 2002) and the effect of a bout of damaging
eccentric work at the onset of a training programme
(Folland, Chong, Copeman, & Jones, 2001), but
these variables were not found to significantly
influence strength gains.
Studies that have employed isometric contractions
have often reported large and rapid increases in
strength [40% in 8 week s (Young, McDonagh, &
Davies, 1985); 25 – 54% in 5 weeks (Thepaut-
Mathieu, Hoecks, & Maton, 1988); 30% in 5 weeks
(Lindh, 1979); 27% in 6 weeks (Weir, Housh, Weir,
& Johnson, 1995)], which could suggest that this
type of training is more effective than conventional
dynamic training. A limitation of isometric training is
that it produces highly length-specific adaptations
with considerable strengt h increases at the training
angle, but with little transfer to other muscle lengths
(Kitai & Sale, 1989; Lindh, 1979; Thepaut-Mathieu
et al., 1988; Weir et al., 1995). In contrast, dynamic
training results in smaller strength increases through-
out the range of the training movement (Graves,
Pollock, Jones, Colvin, & Leggett, 1989).
Although there has been considerable attention to
different types of muscle contractions in resistance
training, few researchers have compared isometri c
and dynamic contractions. Duchateau and Hainaut
(1984) compared maximum isometric contractions
(10 6 5 s duration) with rapid dynamic contractions
(100 at 30 – 40% maximum isometric force). They
found clear evidence for training specificity effects,
with maximum isometric training increasing force
production at high loads, and rapid dynamic
Correspondence: J. P. Folland, School of Sport and Exercise Sciences, Loughborough University, Loughborough LE11 3TU, UK.
E-mail: j.p.folland@lboro.ac.uk
Journal of Sports Sciences, August 2005; 23(8): 817 – 824
ISSN 0264-0414 print/ISSN 1466-447X online ª 2005 Taylor & Francis Group Ltd
DOI: 10.1080/02640410400021783
contractions increasing velocity with low loads.
However, the dissimilar level and duration of loading
make direct comparison of the two types of contrac-
tions impossible.
Jones and Rutherford (1987) used similar high
relative loads to compare isometric, concentric and
eccentric contractions. They found significantly
greater increases in isometric strength (measured at
the isometric training angle) after isometric training
(more than twofold) compared with concentric or
eccentric contractions. Given the documented large
and highly angle-specific adap tations to isometric
training detailed above, this finding is not surprising.
However, Jones and Rutherford also found evidence
for a greater magnitude and duration of muscle
activation during isometric work that may have
accounted for the large isometric strength gains with
this type of training.
In terms of dynamic/isokine tic strength measures,
we are unaware of any studies that have contrasted
isometric and dynamic training at similar relative
loads, perhaps because of the highly ang le-specific
effects that might be expected. Furthermore, no
research to date has contrasted the mode of
contraction (isometric versus dynamic) indepen-
dently from the length specificity adaptation.
Training isometrically at a range of angles might be
more effective than dynamic training, but without
the concentrated angle specificity associated with
isometric training at just one angle. Furthermore,
this approach facilitates a genuine comparison of
dynamic and isometric strength changes following
these two types of contractions.
Kanehisa and Miyashita (1983a) had participants
train isometrically at a range of angles for 8 weeks,
but unfortunately they did not employ a comparison
group that undertook conventional dynamic training.
In the present study, therefore , isometric training at
four different muscle lengths was compared with
conventional dynamic training (lifting and lowering),
using similar relative loads, and assessed by both
isokinetic and isometric strength measures.
Methods
Study desig n
The large individual variation in the response to
strength training (Carey Smith & Rutherford, 1995;
Haakinen, Komi, & Tesch, 1981) makes the
comparison of strength training protocols between
groups of individuals difficult. In contrast, intra-
individual comparisons, where opposite limbs are
trained using different methods, should highlight the
experimental variable. However, crossover effects
that are typically ascribed to neurological adaptations
may confound this type of intra-individual design
(Moritani & deVries, 1979). To minimize the
influence of possible crossover effects, young
healthy, physically active individu als who might have
less scope for changes in learning and coordination
were recruited. Furthermore, to evaluate the capacity
for neurological adaptation, the ability of the
participants to activate the quadriceps muscle was
assessed before training by the twitch interpolation
technique.
Participants
Thirty-three healthy male volunteers (age 21.5 + 2.1
years; body mass 76.5 + 8.6 kg; height 1.81 +
0.06 m; mean + s) completed 9 weeks of knee
extensor strength training. The participants were
recreationally active with no history of knee or thigh
injury, and had not undertaken any leg strength
training during the previous 6 months. They were
recruited from among the staff and students at the
University of Birmingham and gave their informe d
consent to participate. All participants were in-
structed to maintain their habitual level of activity
throughout the study period. The study was
approved by the local ethics committee.
Training
The training consisted of three sessions per week
(Monday, Wednesday and Friday) for 9 weeks and
every training session was supervised. The partici-
pants trained the quadriceps femoris muscle group
unilaterally. One leg of each participants was
randomly assigned to dynamic training, while the
other leg performed only isometric training. The
order of the training (i.e. isometric or dynamic)
within each training session was randomized. The
training load for both protocols was set at 75% of the
respective maximum lift or force for that training
mode, and the maximum was re-assessed on a
weekly basis.
Isometric training
A standard variable resistance leg extension machine
(Cybex VR2) was adapted for isometric work. A
strain gauge was placed in the tension strap and, after
amplification and digitization, the signal was dis-
played on a computer screen in front of the
participant. This allowed force to be measured with
the training apparatus and provided visual feedback
for each contraction. The participants completed
four sets of 10 repetitions of 2 s duration, with one
set being completed at each of four angles of knee
flexion: 0.87, 1.22, 1.57 and 1.92 rad (508,708,908
and 1108). There was 2 s res t between contractions
and 2 min rest between each set. During each
818 J. P. Folland et al.
training session, the sets (angles) were completed in
a different random order.
Dynamic training
Weights were lifted and lowered for four sets of 10
repetitions with a variable resistance leg extension
machine (Cybex, VR2). The participants were
instructed to take 1 s to lift and 1 s to lower each
repetition, through a range of 2.09 to 0.52 rad (1208
to 308 ), equating to *1.57 rad × s
71
, with a short
pause between lifts and 2 min rest between sets.
The variation in loading throughout the range of
motion with the training machine was also assessed.
Using a hand-h eld digital force transducer (Penny
and Giles, Transducers, Christchurch, UK) placed
perpendicular to the lever arm, the force required to
hold a constant load (10 kg) stationary at different
knee flexion angles (0.87, 1.22, 1.57 and 1.92 rad)
was recorded.
Strength testing
Maximum quadriceps strength of each leg was
assessed pre and post training. Pre-training strength
was measured on three occasions, each 1 week apart.
Post-training strength was measured twice, 3 and 5
days after the last training session. The average
values from the pre- and post-training measurements
were compared to evaluate the gains in strength.
Three different types of strength measurements were
made on each test occasion, which lasted approxi-
mately 40 min. There wa s 15 min rest between the
dynamometer measurements (angle – torque and
isokinetic) and isometric strength at 1.57 rad.
Two sets of measurements were made using a
Cybex Norm isokinetic dynamometer (Lumex Inc.,
Ronkokama, NY, USA). The axis of the knee joint
was aligned with the centre of rotation of the
dynamometer arm, and the lower leg was strapped
to the lever arm at the ankle. The participants were
restrained at the waist, shoulders and the distal part
of the thigh, and the backrest was set at 1.74 rad
(1008) from the horizontal base of the seat.
Angle – torque relationship
Isometric strength was al so measured at four angles
of knee flexion [0.87, 1.22, 1.57, and 1.92 rad (508,
708,908 and 1108)] using the dynamometer. The
angles were selected in a random order for each
participant and the order was maintained on
successive test occasions . The participants attempted
two maximal voluntary contractions of 3 s duration
at each angle, with 20 s between each contraction
and at least 30 s rest between each angle. During
each maximal voluntary contraction, the participants
received direct visual feedback of the force signal as
well as verbal encouragement.
Isokinetic strength
Knee extension strength was measured at three
velocities, 0.79, 2.62 and 5.24 rad × s
71
(45, 150
and 3008 × s
71
). The participants performed three
practice trials, before three maximal efforts were
recorded at each velocity. There was 30 s rest
between each velocity and the highe st peak torque
from the three trials was recorded.
Muscle activation and isometric strength at 1.57 rad
Measurements of isometric strength at 1.57 rad (908)
were duplicated with a conventional isometric
strength testing chair (Parker, Round, Sacco, &
Jones, 1990). This system affords measurement of
muscle activation as well as being highly reliable
(over the three baseline tests the coefficient of
variation was 3.5% versus 6.9% for the Cybex
dynamometer at the same angle).
Force was measured using a calibrated U-shaped
aluminium strain gauge (Jones & Park er, 1989) with
a linear response up to 1000 N. The participants
performed three maximal voluntary contracti ons of
the leg extensors with at least 30 s between each.
During each maximal voluntary contraction, the
participants received direct visual feedback of the
force signal as well as verbal encouragement.
On one of the pre-testing occasions, electrically
stimulated twitches were superimposed on three
maximal voluntary contractions to estimate the level
of quadriceps activation (Rutherford, Jones, & New-
ham, 1986). Two conducting rubb er electrodes
(*100 cm
2
), with a coating of conducting gel, were
applied proximally and distally to the anterior surface
of the thigh. A CED-1401 (Cambridge Electronic
Design Ltd, UK) triggered the electrical stimuli
(pulse width 50 ms, up to 200 V; Digitimer DS7,
UK) at a frequency of 1.25 Hz and twitch magnitude
was manipulated by changing the current (range 28 –
50 mA). The size of the twitches during the
voluntary contractions was compared with that at
rest before the contraction to calculate the level of
muscle activation.
Statistical analyses
The data from both dynamometers were expressed
as absolute and relative changes in strength. A three-
way repeated-measures analysis of variance (ANO-
VA, SPSS v11) was performed on the absolute
isometric (time 6 angle 6 training) and absolute
isokinetic (time 6 velocity 6 training) data pro-
duced with the Cybex dynamometer. Re lative values
Isometric versus dynamic strength training 819
from this dynamometer were compared with a two-
way ANOVA for isometric (angle 6 training) and
isokinetic measurements (velocity 6 training). In
each case, Mauchly’s test of sphericity was used to
determine whether the assumption of sphericity was
violated by the data. Where this did occur, the Huyn-
Feldt correction was appli ed. When differences were
found by ANOVA, Tukey’s HSD test was used as a
post-hoc test to ascertain whe re the difference lay.
The data recorded from the conventional strength
chair were evaluated for significance differences
between the training protocols using paired Stu-
dent’s t-tests. The results are expressed as the
mean + standard error of the mean unless stated
otherwise and statistical significance was set at P
5 0.05.
Results
Angle – torque relationship
The angle – torque relationships of the isometrically
trained and dynamically trained legs were very
similar at the start of the study (Figure 1a). Strength
training, irrespective of type, significantly increased
the isometric strength of the participants at a range of
angles (F
1,32
= 115.9, P 5 0.01). The improvement
in absolute isometric strength was significantly
affected by the type of training, with greater
improvements associated with isometric training
(F
1,32
= 9.0, P 5 0.01). Relati ve gains in isometric
strength were also greater for the isometrically
trained than the dynamically trained leg
(F
1,32
= 7.2, P 5 0.01) (Figure 1b). The percentage
gains in isometric strength varied significantly
depending on the measurement angle (F
3,96
= 11.3,
P 5 0.01), with post-hoc analysis revealing gains at
1.57 rad to be greater than at 0.87 and 1.92 rad (P
5 0.01) and gains at 1.22 rad to be greater than at
1.92 rad (P 5 0.01).
The normalized angle – force relationship for the
training machine, specifically the force required to
hold a constant load stationary at different angles of
knee flexion (equivalent to isometric force), dis-
played only a small variation throughout the range of
movement (5 10% variation) (Figure 2). In contrast,
the isometric angle – torque relationship demon-
strated the ability of the quadriceps muscle to vary
by 42% throughout the same range (Figure 1a).
Isokinetic strength
Prior to training, isokinetic strength of the isome-
trically and dynamically trained legs was very similar
at the three measured velocities of 0.78 rad × s
71
(241.7 + 6.9 vs. 240.2 + 6.3 Nm resp ectively),
trained leg, 240.2 + 6.3 N × m), 2.62 rad × s
71
(isometrically trained leg, 175.1 + 4.5 N × m; dyna-
mically trained leg, 173.7 + 4.6 N × m) and 5.24
rad × s
71
(isometrically trained leg, 117.8 +
Figure 1. (a) The angle – torque relationship, before (open
symbols) and after (closed symbols) isometric (triangles) and
dynamic (circles) training. (b) Percentage increase in isometric
strength at each angle after isometric (shaded bars) and dynamic
(open bars) training (mean + s
x
).
Figure 2. The normalized force required to hold a constant load
stationary at different positions throughout the range of movement
with the Cybex VR2 leg extension machine. Force was normalized
to peak force at 1.22 rad.
820 J. P. Folland et al.
2.62 rad × s
71
(175.1 + 4.5 vs. 173.7 + 4.6 Nm) and
5.24 rad × s
71
(117.8 + 3.1 vs. 116.3 + 3.3 Nm).
Resistance training significantly increased absolute
isokinetic strength at this range of velocities, irre-
spective of the type of training undertaken
(F
1,32
= 70.6, P 5 0.01). The improvements in
absolute isokinetic strength with training were not
affected by the type of training (F
1,32
= 0.02,
P = 0.99). Neither was there an interactive effect of
the type of training on the absolute isokinetic
strength at particular velocities (F
2,64
= 1.96,
P = 0.16).
The type of training did not significantly influence
the relative increases in isokinetic strength per se
(F
1,31
= 0.05, P = 0.83) (Fig ure 3), but did interact
significantly with the gains in isokinetic strength at
different velocities (F
2,62
= 3.6, P = 0.03). However,
post-hoc tests revealed no significant differen ce
between isometric and dynamic training at any
specific velocity.
The relative improvements in isokinetic strength
were significantly influenced by the measurement
velocity (F
2,62
= 5.6, P 5 0.01), and post-hoc analy-
sis revealed greater strength gains at 0.79 rad × s
71
than at 5.24 rad × s
71
, irrespective of the type of
training (P 5 0.05).
Muscle activation and isometric strength at 1.57 rad
During the baseline measurements, the participants
were able to achieve 97.2 + 2.1% (mean + s) of full
activation during maximum isometric contracti ons,
as measu red by the twitch interpolation technique.
Before training, the isometrically and dynamically
trained legs were very similar (647.8 + 17.3 N and
652.8 + 17.7 N. respectively). Both types of training
elicited significant increases in absolute strength (P
5 0.001). There was a significantly greater increase
in strength for the isometrically trained leg than the
dynamically trained leg (15.2 + 1.3% and
11.5 + 1.0%, respectively; P 5 0.001).
Discussion
Both types of resistance training resulted in sign ifi-
cant improvements in isometric and isokinetic
strength. Isometric training at four joint angles did
not result in the highly angle-specific adaptations that
have been reported for isometric training at just one
position (Kitai & Sale, 1989; Lindh, 1979; Thepaut-
Mathieu et al., 1988; Weir et al., 1995). However,
training isometrically produced significantly greater
gains in isometric strength across a range of angles
(assessed with two dynamometers) than training
dynamically. In contrast, both types of training
resulted in similar gains in isokinetic (dynamic)
strength.
The current study was a first attempt to make a
direct compa rison between isometric and dynamic
training contractions, while attempting to negate the
confounding factors of differences in relative loading
(magnitude and duration), and the angle specificity
effect of training isometrically at just one angle. The
experiment was designed so that both training
protocols had an equal duration of tension at the
same relative load. However, there are some issues in
the control of these parameters that could have
influenced the results.
First, even a small discrepancy in the duration of
loading may have an accumulative effect upon
strength gains, as noted by Jones and Rutherford
(1987). Although every training session of each
participant was strictly supervised, during dynamic
training there can be a natural tendency to lift and
lower the weight at a rate of greater than 2 s per lift.
The authors are confide nt that for voluntary training
the duration was matched as closely as possible.
Second, there are a number of aspects to be
considered when comparing the loading of the two
protocols. The intention of equal relative load ing for
the two protocols is clearly complicated when one
considers that the dynamic training involved lifting
(concentric) and lowering (eccentric) phases that
have diverse force capabilities. In the current study,
the intention was to match the relative loading for the
lifting (concentric) phase of the dynamic training
with the isometric training. As maximum isometric
strength is greater than concentric strength, this
matched relative loading accepted a discrepancy in
the absolute level of loading. The dynamic training
involved an average velocity of 1.57 rad × s
71
and the
load was set relative to 1-RM, which was presumably
determined by concentric lifting strength. Force –
velocity data from similar subjects (Folland et al.,
2002) demonstrated peak concentric torque at 1.57
rad × s
71
was ~75% of peak isometric torque at the
Figure 3. Percentage increase in isokinetic peak torque at three
angular velocities after isometric (shaded bars) and dynamic
training (open bars) (mean + s
x
).
Isometric versus dynamic strength training 821
same angle (*758 of knee flexion). Therefore, this
discrepancy in absolute loading (i.e. 33% greater
for the isometrically trained than the dynamically
trained leg) was accepted from the onset of the
study to contrast equal relative loading. As there
has been little work comparing these different types
of contractions, it is unclear whether it is absolute
or relative loading that is the critical parameter in
the training response, and in the current study
which would pr ovide the more valid comparison.
An alternative methodology would be to attempt to
match absolute torque in the two training proto-
cols. This approach would clearly negate matched
relative loading and might necessitate sub-optimal
isometric loading in order to balance absolute
torques.
Furthermore, due to the mechanics of the exercise
machine used for the dynamic training, the actual
dynamic training load seems certain to have been
lower than intended for much of the range of
movement. In the current study, we found that a
modern well-en gineered resistance training machine
(Cybex, VR2), with a var iable cam, did not
adequately match the angle – torque relationship of
the quadriceps muscle of the participants. It is our
belief that this is commonly the case even with
modern resistance training apparatus. The angle –
force relationship for the training machine was very
flat ( 5 10% variation) (Figure 2) in comparison
with the muscle’s ability (Figure 1a), whi ch varied
substantially (42%) throughout the same range of
movement. The contrast of these two curves suggests
that the greatest relative loading will be at the
periphery of the range of m otion – particularly at
long muscle lengths where the muscle is at its
weakest. It is therefore not surprising that the
commonly observed ‘‘sticking point’’ limiting a lift
is at the beginning of the movement (long muscle
lengths, 5 1.92 rad), especially when one considers
that the inertia of the load must also be overcome at
this point. If the maximum lift (1-RM) was limited
by strength at this point, then the prescribed relative
training load (75% 1-RM) is likely to only have
provided the desired loading at this point, with less
than the prescribed training load during the remain-
der of the movement. For example, as the movement
progressed to an angle of 1.22 rad (708), in contrast
to 1.92 rad (1108), there is a disproportionate
increase in the muscle’s ability compared with the
small additional torque required at this angle. It can
be estimated that at 1.22 rad the same lift would
equate to only 58% of maximum concentric torque,
rather than the prescribed 75%. This implies that the
dynamic training load may have varied between 58
and 75% of isometric training torque, according to
the angle under consideration, and implies the
isometric training load was 33 – 75% greater than
the dynamic training load. This clearly represents a
substantial discrepancy.
In an attempt to compare isometric and dynamic
loading independent of angle specificity, we tried to
match the relative loading of the two protocols. In
retrospect, due primarily to the surprisingly flat
nature of the angle – force relationship of the training
machine, this was not achieved, and this accentuated
the difference in absolute torque of the two training
protocols. Future work would benefit from a more
uniform relative loading throughout the range of
motion for the dynamic training so as to accurately
equate the relative loading.
The overall findings from the two dynamometers
used for isometric measurement were similar (Cybex
Norm and conventional strength chair: significantly
greater isometric strength gains with isometric
training), but in terms of the magnitude of the gains
in isometric strength at 1.57 rad, there was a clear
discrepancy between them (conventional strength
chair: dynamically trained vs. isometrically trained
leg, 11.5% vs. 15.2 %; Cybex Norm: dynamically
trained vs. isometrically trained leg, 20.0% vs.
21.9%). It is not clear why there was such a
difference in the magnitude of recorded strength
gains (1.4 – 1.7-fold greater for the Cybex dynam-
ometer). It may be partially attributed to the lower
reliability of the Cybex (coefficient of variation: 6.9%
vs. 3.5%). Most commercial dynamometers are
designed primaril y for rehabilitation and their pad-
ding reduces the reproducibility of positioning the
participant and causes greater compliance within the
measurement system. Additionally in the current
study, only two maximal voluntary contractions were
attempted at each angle with the Cybex, as opposed
to three with the conventional strength chair.
However, it is difficult to see how any difference in
reliability might affect the magnitude of the strength
changes.
To remove the concentrated angle-specific effects
of isometric training at just one angle, four distinct
yet contiguous isometric angles were selected (0.87,
1.22, 1.57 and 1.92 rad). This more diverse
isometric training employed in the present study
did not produce strength gains that were as large as
those reported for isometric training at just one angle
[e.g. 35% after 12 weeks of training (Jones &
Rutherford, 1987)]. This was not surprising con-
sidering that only a quarter of the training stimulus in
the present study was specific to any given angle.
The significantly greater isometric strength gains
with isometric training, compared with dynamic
training, could be attributed to different factors.
One possibility is a residual angle specificity effect.
Although the current isometric training was divided
over four angles, considering the potent angle
specificity effect observed with isometric training at
822 J. P. Folland et al.
just one angle, there may still have been a residual
angle specificity effect. In particular, greater gains in
isometric strength at the training angles, but smaller
gains at other angles. In contrast, the dynamic
training involved a larger range of motion (dynamic
vs. isometric: 0.52 – 2.09 vs. 0.87 – 1.92 rad) and a
more diffuse training stimulus. Unfortunately, the
current study did no t include isometric strength
measurement at angles between or outside of the
training angles, but this would be strongly advised in
future research.
The greater gains in isometric strength with
isometric training could be due to a contractile
mode specificity effect, with isometric training
producing neurophysiological adaptations specific
to isometric contractions. Although there is strong
evidence for a contractile mode spe cificity effect
when contrasting concentric and eccentric training
(Hortobagyi et al., 1996), independent of an angle
specificity effect, the authors are not awar e of any
evidence for a contractile mode specificity discre-
pancy between isometric and concentric strength.
Finally, and perhaps most likely, the higher
absolute torques associated with isometric training
(estimated as 33 – 75% higher) may account for the
greater isometric strength gains observed. This
appears to be a substantial difference, particularly
as the level of loading is considered critical to the
training response (Atha, 1981; McDonagh & Davies,
1984), and therefore seems a probable explanation
for the greater isometric strength gains with isometric
training.
The significantly greater strength gains at the mid-
range angles (1.57 and 1.22 rad), irrespective of the
type of training, was an unexpected finding. In terms
of isometric training, it could be hypothesized that
there might be transfer of strength gains at one
position to adjacent angles/positions. After 6 weeks
of isometric training at one angle, Weir et al. (1995)
found significant increases in strength up to 0.52 rad
(308) from the training angle. If this were the case in
the present study, the mid-range angles would
exhibit the greatest strength gains as they would
receive tran sfer effects from both adjacent shorter
and longer muscle lengths. The greater gains in
isometric strength at mid-range angles with dynamic
training is contrary to our previous findings (Folland
et al ., 2002) as well as the proposed rationale that the
highest relative loading occ urred at long muscle
lengths. Our earlier stu dy employed a similar
dynamic training machine, but found significant
increases in isometric strength only at the longer
muscle lengths. The reason for these contradictory
findings is unclear.
Overall, the increases in isokinetic strength were
fairly similar for isometric and dynamic resistance
training. There was no effect of the different types of
training upon isokine tic strength gains per se,orat
any specific velocity. However, from Figure 3 there
appears to be a steeper drop-off in strength gains at
higher velocities for the isometrically trained than the
dynamically trained leg, and the pattern of isokinetic
strength gains across the three velocities was sig-
nificantly different according to the type of training.
This is in agreement with the literature, which
indicates a degree of velocity specificity in strength
training (Caiozzo, Perrine, & Edgerton, 1981; Coyle
et al., 1981; Kanehisa & Miyashita, 1983b; Moffroid
& Whipple, 1970). The fact that only half of the
dynamic training involved concentric activity may
also have confounded the chances of finding a
velocity-specific effect in the current study. While it
is dynamic lifting and lowering that is the widely
practised form of resistance training, a comparison of
purely concentric and isometric work would provide
a more interesting neurophysiological comparison.
Isokinetic strength gains were significantly greater
at 0.79 rad × s
71
than at 5.24 rad × s
71
, irrespective of
the type of training. The training velocities for both
types of training (isometric, 0; dynamic, 1.57
rad × s
71
) were closest to the slowest isokinetic test
velocity of 0.79 rad × s
71
, and most distinct from the
fastest test velocity of 5.24 rad × s
71
. This provides
further evidence for a velocity specificity effect.
In conclusion, training isometrically at four angles
produced significantly greater gains in isometric
strength across a range of angles (assessed with two
dynamometers), but similar gains in isokinetic
(dynamic) strength in comparison to dynamic train-
ing. The greater isometric strength gains could be
due to a residual angle specific ity effect or, perhaps
more likely, the greater absolute torque involved with
isometric training.
References
Atha, J. (1981). Strengthening muscle. Exercise and Sport Sciences
Reviews, 9, 1 – 73.
Caiozzo, V. J., Perrine, J. R., & Edgerton, V. R. (1981). Training
induced alterations of the in-vivo force – velocity relationship of
human muscle. Journal of Applied Physiology, 51, 750 – 754.
Carey Smith, R., & Rutherford, O. M. (1995). The role of
metabolites in strength training, I. A comparison of eccentric
and concentric contractions. European Journal of Applied
Physiology, 71, 332 – 336.
Coyle, E. F., Feiring, D. C., Rotkis, T. C., Cote, R. W., III, Roby,
F. B., Lee, W. et al. (1981). Specificity of power improvements
through slow and fast isokinetic training. Journal of Applied
Physiology, 51, 1437 – 1442.
Duchateau, J., & Hainaut, K. (1984). Isometric or dynamic
training: Differential effects on mechanical properties of a
human muscle. Journal of Applied Physiology, 56, 296 – 301.
Folland, J. P., Chong, J. C., Copeman, E. M., & Jones, D. A.
(2001). Acute muscle damage as a stimulus for training-
induced gains in strength. Medicine and Science in Sports and
Exercise, 33, 1200 – 1205.
Isometric versus dynamic strength training 823
Folland, J. P., Irish, C. S., Roberts, J. C., Tarr, J. E., & Jones, D.
A. (2002). Fatigue as a stimulus for strength gains during
resistance training. British Journal of Sports Medicine, 36, 370 –
374.
Graves, J. E., Pollock, M. L., Jones, A. E., Colvin, A. B., &
Leggett, S. H. (1989). Specificity of limited range of motion
variable resistance training. Medicine and Science in Sports and
Exercise, 21, 84 – 89.
Haakinen, K., Komi, P. V., & Tesch, P. A. (1981). Effect of
combined concentric and eccentric strength training and
detraining on force – time, muscle fibre, and metabolic char-
acteristics of leg extensor muscles. Scandinavian Journal of
Sports Science, 3, 50 – 58.
Hortobagyi, T., Hill, J. P., Houmard, J. A., Fraser, D. D.,
Lambert, N. J., & Israel, R. G. (1996). Adaptive responses to
muscle lengthening and shortening in humans. Journal of
Applied Physiology, 80, 765 – 772.
Jones, D. A., & Parker, D. F. (1989). Development of a portable
strain gauge to measure human muscle isometric strength.
Journal of Physiology, 145, 11P.
Jones, D. A., & Rutherford, O. M. (1987). Human muscle
strength training: The effects of three different training regimes
and the nature of the resultant changes. Journal of Physiology,
391, 1 – 11.
Kanehisa, H., & Miyashita, M. (1983a). Effect of isometric and
isokinetic muscle training on static strength and dynamic
power. European Journal of Applied Physiology, 50, 365 – 371.
Kanehisa, H., & Miyashita, M. (1983b). Specificity of velocity in
strength training. European Journal of Applied Physiology, 52,
104 – 106.
Kitai, T. A., & Sale, D. G. (1989). Specificity of joint angle in
isometric training. European Journal of Applied Physiology, 58,
744 – 748.
Lindh, M. (1979). Increase of muscle strength from isometric
quadriceps exercise at different knee angles. Scandinavian
Journal of Rehabilitative Medicine, 11, 33 – 36.
McDonagh, M. J. N., & Davies, C. T. M. (1984). Adaptive
response of mammalian skeletal muscle to exercise with high
loads. European Journal of Applied Physiology, 52, 139 – 155.
Moffroid, M. T., & Whipple, R. H. (1970). Specificity of speed of
exercise. Physical Therapy, 50, 1693 – 1699.
Moritani, T., & deVries, H. A. (1979). Neural factors versus
hypertrophy in the time course of muscle strength gain.
American Journal of Physical Medicine, 58, 115 – 130.
Parker, D. F., Round, J. M., Sacco, P., & Jones, D. A. (1990). A
cross-sectional survey of upper and lower limb strength in boys
and girls during childhood and adolescence. Annuals of Human
Biology, 17, 199 – 211.
Rutherford, O. M., Jones, D. A., & Newham, D. J. (1986).
Clinical and experimental application of the twitch interpola-
tion technique for the study of human muscle activation.
Journal of Neurology, Neurosurgery and Psychiatry, 49, 1288 –
1291.
Thepaut-Mathieu, C., Hoecke, J., & Maton, B. (1988). Myoelec-
trical and mechanical changes linked to length specificity during
isometric training. Journal of Applied Physiology, 64, 1500 –
1555.
Weir, J. P., Housh, T. J., Weir, L. L., & Johnson, G. O. (1995).
Effects of unilateral isometric strength training on joint angle
specificity and cross-training. European Journal of Applied
Physiology, 70, 337 – 343.
Young, K., McDonagh, M. J., & Davies, C. T. (1985). The effects
of two forms of isometric training on the mechanical properties
of the triceps surae in man. Pflu¨gers Archives, 405, 384 – 388.
824 J. P. Folland et al.